Ghostsign ActionsInstall

Summary

Ingest a project's text or file into Ghostsign embeddings for retrieval‑augmented generation.

Inputs

• projectId (required) — Identifier of the target project in Ghostsign.
• embedSourceMode — Selects the input source: 'Manual Text Chunk' or 'Storage Object'.
• embedManualText — Plain text to embed; required when embedSourceMode is Manual Text Chunk.
• embedStoragePath — Fully qualified path in the ghostsign-context bucket (e.g., orgId/doc.pdf) for storage mode.
• embedContentType — MIME type hint for the stored file if it cannot be inferred from the extension.
• embedNoteLabel — Optional label stored as metadata on the created embedding batch.

Output shape

a list of response objects, one per input item, containing the result JSON from the Ghostsign extract‑embed endpoint (typically includes an embed_batch_id).

The exact payload depends on Ghostsign; fields such as embed_batch_id may be present. Errors are handled via continue‑on‑fail and returned as error objects.

Examples

Example 1: Add a manually typed description into the embeddings store.

Set embedSourceMode to Manual Text Chunk, provide Project ID, and paste text into embedManualText.

Example 2: Ingest a PDF stored in the ghostsign-context bucket.

Choose Storage Object as embedSourceMode, fill embedStoragePath with the bucket URI, optionally set embedContentType, and optionally add embedNoteLabel.

Example 3: Create an embedding with associated note metadata.

Use Manual Text Chunk mode, fill Project ID and embedManualText, then set embedNoteLabel to tag the batch.

Links

• n8n community nodes installation guide